Telco AI Deep Dive: Tuning LLMs for telco-specific gen AI

3890

In order to effectively benefit from telco AI solutions, operators need to first focus on building a robust data platform

For telco AI use cases to materialize in a beneficial manner, operators first need to understand and organize the data they have which is, generally speaking, very personal, highly contextualized and historically underutilized. In an interview with RCR Wireless News, AWS Chief Technologist for for Telecom and Edge Cloud Ishwar Parulkar discussed the importance of getting data out of organizational silos and into a data platform that can then be used to optimize existing large language models (LLMs) that, in turn, power generative AI (gen AI).

“Telcos have a lot of data,” he said. “They, I think, are recognizing that it all starts with managing and…using the data properly. What the gen AI revolution or transformation is doing is making them think very hard at that, and we see a lot of efforts around really starting looking at the basics. This is going to be the trigger that’s going to really make them become more data-centric.”

In the context of the multiple technological and organizational shifts operators are working towards—going to 5G Standalone and embracing cloud-native, beefing up enterprise revenues with private 5G and mobile edge computing solutions, and automating what can be automated—gen AI could be the catalyst that helps otherwise stagnate financial performance be jumpstarted and (hopefully) lead to growth.

“Telcos are uniquely positioned because they have this new network capability with the edge,” Parulkar said. “They can really start looking at their business in a very different manner. “

Click here to see other interviews in the ongoing Telco AI Deep Dive series.